Research Information System
31st International Conference on Artificial Life and Robotics, ICAROB 2026, Oita, Japan, 29 January - 01 February 2026, pp.211-215, (Full Text)
Agriculture is losing its popularity in Türkiye because of outdated techniques. Object detection is essential for enhancing productivity, harvesting, and yield estimation for farmers. In this study, an autonomous tomato detector utilizing an image processing approach in Python is developed, leveraging a dataset comprising diverse tomato types under various environmental and lighting conditions to enhance the application. This approach utilized traditional image processing techniques, including masking and filtering. To overcome the overlap problem that occurred in crowded tomato images, the watershed split algorithm was used to separate overlapping tomatoes to improve detection precision. The application can effectively distinguish tomatoes in the field to be harvested. Removing the overlap problem significantly raises the detection performance. The outcome of this study suggests that computerbased techniques can help agriculture while lowering labour costs.